TUTCRIS - Tampereen teknillinen yliopisto

TUTCRIS

Feature synthesis for image classification and retrieval via one-against-all perceptrons

Tutkimustuotosvertaisarvioitu

Standard

Feature synthesis for image classification and retrieval via one-against-all perceptrons. / Raitoharju, Jenni; Kiranyaz, Serkan; Gabbouj, Moncef.

julkaisussa: Neural Computing and Applications, Vuosikerta 29, Nro 4, 02.2018, s. 943–957.

Tutkimustuotosvertaisarvioitu

Harvard

APA

Vancouver

Author

Raitoharju, Jenni ; Kiranyaz, Serkan ; Gabbouj, Moncef. / Feature synthesis for image classification and retrieval via one-against-all perceptrons. Julkaisussa: Neural Computing and Applications. 2018 ; Vuosikerta 29, Nro 4. Sivut 943–957.

Bibtex - Lataa

@article{02962792d79e48a99cf391a850c17523,
title = "Feature synthesis for image classification and retrieval via one-against-all perceptrons",
abstract = "Most existing content-based image retrieval and classification systems rely on low-level features which are automatically extracted from images. However, often these features lack the discrimination power needed for accurate description of the image content, and hence, they may lead to a poor retrieval or classification performance. We propose a novel technique to improve low-level features which uses parallel one-against-all perceptrons to synthesize new features with a higher discrimination power which in turn leads to improved classification and retrieval results. The proposed method can be applied on any database and low-level features as long as some ground-truth information is available. The main merits of the proposed technique are its simplicity and faster computation compared to existing feature synthesis methods. Extensive simulation results show a significant improvement in the features’ discrimination power.",
keywords = "Content-based image retrieval and classification, Feature synthesis, Multi-dimensional particle swarm optimization, Multi-layer perceptrons",
author = "Jenni Raitoharju and Serkan Kiranyaz and Moncef Gabbouj",
note = "EXT={"}Kiranyaz, Serkan{"}",
year = "2018",
month = "2",
doi = "10.1007/s00521-016-2504-4",
language = "English",
volume = "29",
pages = "943–957",
journal = "Neural Computing and Applications",
issn = "0941-0643",
publisher = "Springer Verlag",
number = "4",

}

RIS (suitable for import to EndNote) - Lataa

TY - JOUR

T1 - Feature synthesis for image classification and retrieval via one-against-all perceptrons

AU - Raitoharju, Jenni

AU - Kiranyaz, Serkan

AU - Gabbouj, Moncef

N1 - EXT="Kiranyaz, Serkan"

PY - 2018/2

Y1 - 2018/2

N2 - Most existing content-based image retrieval and classification systems rely on low-level features which are automatically extracted from images. However, often these features lack the discrimination power needed for accurate description of the image content, and hence, they may lead to a poor retrieval or classification performance. We propose a novel technique to improve low-level features which uses parallel one-against-all perceptrons to synthesize new features with a higher discrimination power which in turn leads to improved classification and retrieval results. The proposed method can be applied on any database and low-level features as long as some ground-truth information is available. The main merits of the proposed technique are its simplicity and faster computation compared to existing feature synthesis methods. Extensive simulation results show a significant improvement in the features’ discrimination power.

AB - Most existing content-based image retrieval and classification systems rely on low-level features which are automatically extracted from images. However, often these features lack the discrimination power needed for accurate description of the image content, and hence, they may lead to a poor retrieval or classification performance. We propose a novel technique to improve low-level features which uses parallel one-against-all perceptrons to synthesize new features with a higher discrimination power which in turn leads to improved classification and retrieval results. The proposed method can be applied on any database and low-level features as long as some ground-truth information is available. The main merits of the proposed technique are its simplicity and faster computation compared to existing feature synthesis methods. Extensive simulation results show a significant improvement in the features’ discrimination power.

KW - Content-based image retrieval and classification

KW - Feature synthesis

KW - Multi-dimensional particle swarm optimization

KW - Multi-layer perceptrons

U2 - 10.1007/s00521-016-2504-4

DO - 10.1007/s00521-016-2504-4

M3 - Article

VL - 29

SP - 943

EP - 957

JO - Neural Computing and Applications

JF - Neural Computing and Applications

SN - 0941-0643

IS - 4

ER -